HSR&D Home » Research » IIR 12-358 – HSR&D Study
Improving Surgical Quality: Risks and Impact of Readmission
Melanie S Morris, MD
Birmingham VA Medical Center, Birmingham, AL
Mary Hawn MD MPH BS
VA Palo Alto Health Care System, Palo Alto, CA
Palo Alto, CA
Funding Period: December 2014 - March 2018
Hospital readmissions have been targeted as a hospital quality measure. Readmissions can increase both costs and resource utilization and are associated with poorer patient outcomes. While much research on readmissions has been done in the medical patient population, there has been little study of reasons for readmission in the surgical patient population. It will be important to identify which patients are at high risk for readmission after surgery and to understand whether a readmission is potentially preventable, represents a quality of care issue, or indicates failure of the care transition plan. By incorporating the contributions of patient comorbidity, self-efficacy, caregiver status, procedure complexity, and system factors on readmissions we can develop a risk prediction tool to identify those patients at highest risk.
1. Evaluate the contribution of patient, procedure, post-operative complication and system factors on readmission within 30 days of hospital discharge following surgery, and use these data to develop and validate a readmission risk prediction tool that can be used real-time, develop a classification of readmission reasons, and explore processes of care linked with readmission.
2. Assess potential patient factors not currently collected by VASQIP at discharge and determine their association with readmission.
3. Rank reasons for readmission categories developed from Aims 1 and 2 as potentially preventable and appropriateness as a measure of surgical quality.
A detailed analysis of surgical readmissions using VA Surgical Quality Improvement Program (VASQIP) data linked with clinical and administrative data containing information on index admission vital signs, pain scores and laboratory data will be used to develop a readmission risk prediction tool. Concurrent with these analyses, we will perform prospective data collection on patient health literacy and caregiver status along with prospective evaluation of the readmission risk prediction tool. This will allow us to characterize clinical and social determinants of readmission to further refine our risk prediction tool. Using a Delphi process, we will develop a readmission classification system that categorizes readmission reasons as potentially preventable and whether they reflect quality of surgical care.
Final analyses addressing Aim 1 have been completed with results published in Annals of Surgery. The analytic study sample included 237,441 surgeries: 43% orthopedic, 39% general and 18% vascular. Overall 30-day unplanned readmission rate was 11.4%, differing by surgical specialty (vascular 15.4%, general 12.9%, and orthopedic 7.6%, p<0.001). The most common readmission reasons were wound complications (30.7%), GI (16.1%), bleeding (4.9%), and fluid/electrolyte (7.5%) complications. Models using information available at the time of discharge explained 10.3% of the variability in readmissions. Of these, preoperative patient-level factors contributed the most to predictive models (R² 7.0% [c-statistic 0.67]); prediction was improved by inclusion of intraoperative (R² 9.0%, c-statistic 0.69) and postoperative variables (R² 10.3%, c-statistic 0.71). Including post-discharge complications improved predictive ability, explaining 19.6% of the variation (R² 19.6%, c-statistic 0.76).
Secondary analyses examining specific contributors to readmission have been published and are currently ongoing.[2-5] In summary, within the same hospital, readmission rates for three surgical specialties were not correlated and there was little correlation between procedures within specialties. Little of the variation in readmissions was attributable to hospital or specialty level factors. These findings suggest that postoperative readmissions are mostly related to patient-level factors as opposed to hospital or specialty effects. In addition, postoperative pain trajectories identify populations at risk for 30-day readmissions and emergency department visits. This relationship does not appear to be mediated by post-discharge complications. Addressing pain control expectations prior to discharge may help reduce surgical readmissions in high pain categories. In secondary analyses addressing frailty, the modified Frailty Index was associated with poor surgical outcomes primarily due to impaired functional status. Efforts to further characterize and optimize potentially modifiable aspects of frailty preoperatively, specifically improving functional status, may improve perioperative outcomes including unplanned readmission. Lastly, early postoperative hyperglycemia was associated with increased readmission but elevated preoperative HbA1c is not. Higher preoperative HbA1c was associated with increased postoperative glucose checks and insulin use, suggesting that heightened postoperative vigilance and a lower threshold to treat hyperglycemia may explain this finding. In addition, our study team completed and published a methods paper detailing the study design in BMC Health Services Research.
A Delphi panel was convened to address Aim 3. The 14 panelists rated readmission reasons over three rounds and 12 participated in the phone call. Wound-related infections, sepsis, urinary tract infections, pneumonia, hemorrhage or hematoma, anemia, ostomy complications, catheter-related bloodstream infection, acute renal failure and other fluid and electrolyte disorders and postoperative venous thromboembolism were all considered related to surgical quality by our panel of experts. Heart problems, gastrointestinal complaints, seizures, stroke, pain, injuries, and mental health were considered unrelated despite their relatively high prevalence among surgical readmissions (40%).
Data collection and follow-up in the prospective arm of the study (Aim 2) concluded in July 30, 2017 with 749 patients enrolled and an overall readmission rate of 17.0% across all sites. Data cleaning and initial reporting is complete. Preliminary analysis shows surgical readmissions are difficult to predict, but health literacy and post-acute care utilization may influence it. Many factors in the retrospective model stayed consistent with the prospective cohort. Prospectively collected data showed lower health literacy is associated with higher readmission rates and being discharged to an inpatient rehabilitation center or skilled nursing facility is associated with lower readmission rates.
The products of this study will have a direct impact on patient care by informing the specifications of a real-time risk prediction tool. In developing standardized predictors and metrics of readmission rates, we will be better positioned to assess the quality of surgical care in VA and improve our tracking and reporting of surgical outcomes. In interpreting our results, we described the implications for interventions with patients and caregivers to reduce postoperative readmissions, as well as informed development of discrete readmission tracking variables for VASQIP.
External Links for this Project
NIH ReporterGrant Number: I01HX001108-01A2
Dimensions for VADimensions for VA is a web-based tool available to VA staff that enables detailed searches of published research and research projects.
If you have VA-Intranet access, click here for more information vaww.hsrd.research.va.gov/dimensions/
VA staff not currently on the VA network can access Dimensions by registering for an account using their VA email address. Search Dimensions for this project
DRA: Health Systems
DRE: Technology Development and Assessment
MeSH Terms: none